JP2004220152A - Information recommendation device, extraction device for information recommendation destination, computer program and computer readable recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide information valid for a user at a proper timing. <P>SOLUTION: This information recommendation device is provided with a storage device for storing purchase history information 5 and purchase cycle information 6 of merchandise of a plurality of users, a function for extracting purchased merchandise change history 6 of the users, a function for deciding whether or not a period when a certain user should change the purchased merchandise arrives from the purchase history information 5 and the purchase cycle information 6, a function for deciding merchandise 8 to be recommended to the certain user based on the purchase change history of the other users when it is decided that the period when the user should change the purchased merchandise arrives and functions 9 and 10 for providing information associated with the decided merchandise to the certain user. <P>COPYRIGHT: (C)2004,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention generally relates to a system for recommending a product to be purchased next to a user or a time when a product should be purchased next based on a purchase history of the user. The present invention particularly provides an information recommendation device that determines information about a product recommended to a user from information about a user's past purchase history of products and information about a purchase cycle, an information recommendation destination extraction device, a program therefor, and a program therefor. The present invention relates to a recording medium on which a program is recorded.
[0002]
[Prior art]
In order to provide useful shopping information to the user from the user's purchase history information, there is a method of using information about a user's purchase history and information about a purchase cycle for a certain product. Patent Document 1 listed below discloses such a technique. In the invention described in Patent Literature 1, it is determined from a user's purchase history information on a certain product whether or not the purchase time of the user on the product has periodicity, and the purchase schedule of the product is determined based on the information on the purchase cycle. A technique of proposing the user to purchase the product near the day is adopted.
[0003]
Further, in the invention described in Patent Document 1, it is possible to provide new product information to a user by referring to the history of another user. That is, by comparing attribute information such as age, income, and purchase history between users, a similar user is selected, and a purchase history of a user similar to a certain user is referred to infer a product likely to be purchased by the user. And suggestions.
[0004]
Here, the comparison of the purchase histories is a comparison regarding the purchased products as shown in FIG. As can be seen from FIG. 17, both the user X and the user Y have purchased the products A, B, and C, and since the purchased products are similar, the users are assumed to be similar users. User X has purchased product P, but user Y has not purchased this product. Similarly, user Y has purchased product Q, but user Y has not purchased product Q. The invention described in Patent Literature 1 is based on the premise that users having a similar group of products purchased in the past have a similar preference for the products. Therefore, in the invention described in Patent Literature 1, the user X recommends a product Q that has not been purchased by the user X and has been purchased by the user Y who is a similar user, and the user Y has not purchased the user Y. In addition, a product P purchased by a user X who is a similar user is recommended.
[0005]
[Patent Document 1]
JP 2001-22826 A
[Problems to be solved by the invention]
However, in the invention disclosed in Patent Document 1, for example, when information about a new product Q is provided to a certain user X, the similarity of purchase histories of products (A, B, C) other than the product Q is similar. , Similar users are extracted, and a recommended product is selected from the extracted purchase history of the products of the similar user. Therefore, there is a problem that it is unclear whether the recommended product is truly useful for the user.
[0006]
Further, in the method disclosed in Patent Document 1 described above, history information referred to for recommendation is managed based on product identification information, price, and the like. For this reason, it is not possible to deal with fluctuations in taste that do not appear in the identification information or the price of the product, and cannot cope with subtle changes in taste of the user.
[0007]
For example, it is assumed that the user changes his eating habits so as to avoid fat. Taking milk as an example, it is assumed that the user will switch from non-adjusted milk that the user has purchased for a long time to low-fat processed milk. In this case, it is better to recommend a product with low fat content for other products. For example, if the user regularly purchased Camembert cheese, he would like to recommend cottage cheese with lower fat. However, when recommending products only from past purchase history, it is difficult to make such a response, and it is difficult to follow changes in user preferences, or it is difficult to recommend products according to changes in the user's lifestyle. There's a problem.
[0008]
On the other hand, there are the following problems from the viewpoint of the user. For example, in the case of milk, in many cases, a product name that indicates low fat at a glance is displayed, such as "low fat milk". Therefore, it is relatively easy for a user to choose a low-fat food. However, not all products generally have such an indication. For example, in the case of cheese, cottage cheese has less fat than other types of cheese, but this is difficult for ordinary users to understand. Therefore, it is difficult for the user to select an appropriate product even if he / she wants to select low-fat cheese. In such a case, it is desirable to provide appropriate information to the user.
[0009]
In addition, the timing of providing new product information to the user is also important. For example, it is effective to provide new information when a user wants new information. For example, it may be effective to provide information on another product when the user is bored with the product.
[0010]
Therefore, an object of the present invention is to provide an information recommendation device capable of providing effective information for a user at an appropriate timing.
[0011]
Another object of the present invention is to provide an information recommendation apparatus that can cope with a small change in user preference and provide information according to the change.
[0012]
[Means for Solving the Problems]
In order to solve the above problem, in the present invention, a purchase history and a purchase cycle are set for each user, and a purchase change history is extracted. From the three pieces of information, it is determined whether or not the purchased product has changed, and in the case of the changing time, a new product is recommended to the user.
[0013]
The time of change can be determined if the purchase interval of one product purchased regularly is long, or if the purchase product is changed regularly for one item, and the other product This is determined when a change in taste is observed due to a change in purchase of the item.
[0014]
Newly recommended products can be recommended based on purchase change histories of other users who have similar purchase histories, recommend another product in the same item as a certain product, or Or making a recommendation based on a change in preference obtained by a change in the purchase of the user.
[0015]
That is, the information recommendation apparatus according to the first aspect of the present invention is an information recommendation apparatus that recommends information suitable for a user based on purchase history information and purchase cycle information of the user regarding products classified for each item. A history storage unit for storing purchase history information and purchase cycle information of a plurality of users; and a change of a product purchased by the user from the purchase history information and purchase cycle information of the user stored in the history storage unit. A purchase change history extracting unit for extracting a history; a determination unit for determining whether or not it is time to change a purchased product for a certain item from the purchase history information and the purchase cycle information; and a determination unit. In response to the user determining that it is time to change the purchased product for a certain item, the purchase change history of another user extracted by the purchase change history extraction means is returned. The basis, including the product determining means for determining a product to be recommended for a user, and information providing means for providing the user with information about the products which is determined by the product determination means.
[0016]
In this case, for example, it is determined from the information of the purchase history, the purchase cycle, and the purchase change history of each user that the user is tired when the purchase interval of a certain purchased product A that is regularly purchased is long. . From the purchase change history of another user having a similar purchase history, it can be seen that this user has changed from product A to product B. In this case, the product B is recommended to the user. Thus, by providing another product B to a user who is tired of a certain product A, it is possible to provide effective information at a good timing for the user.
[0017]
Preferably, the merchandise determination means searches the history storage means to extract a user who has a purchase tendency similar to a certain user's purchase tendency of a certain commodity and changes a purchase commodity of a certain commodity. It includes a user extracting unit and a unit for determining, as a product to be recommended to a certain user, a purchased product obtained by changing a purchased product of a certain item of the user extracted by the similar user extracting unit.
[0018]
More preferably, the similar user extracting means has a history of purchasing the same product as the purchased product of a certain item of a certain user at substantially the same cycle as the purchase cycle of the purchased product of a certain item of a certain user, and Means for extracting a user who has changed the purchased product is included.
[0019]
When a user purchases a product, a user who has a similar behavior is likely to exhibit the same behavior when changing a purchased product. Therefore, such users who have recently changed the purchased product are extracted, and information on the purchased product after the change is presented to the original user. As a result, effective information can be provided to the original user when changing the product.
[0020]
More preferably, the information providing means is a recommendation information storage means for storing information on the product determined by the product determination means, and the information stored in the recommendation information storage means in response to a predetermined information recommendation request. For providing to a certain user.
[0021]
By storing information in advance in the recommendation information storage means, the information can be provided immediately when an information recommendation request is made. Therefore, effective information can be provided with good timing.
[0022]
An information recommendation destination extraction apparatus according to a second aspect of the present invention is an information recommendation apparatus that recommends information suitable for a user from purchase history information and purchase cycle information of the user regarding products classified for each item. A history storage unit for storing purchase history information and purchase cycle information of a certain user; and a change history of a purchased product from the purchase history information and purchase cycle information of a certain user stored in the history storage unit. The purchase change history extracting means for extracting, the determination means for determining whether it is time for the user to change the purchased product for a certain item from the purchase history information and the purchase cycle information, In response to the determination that it is time to change the purchased product for an item, the information for recommending to a user information about the user's unpurchased products within the same item is And an input item recommendation means.
[0023]
In this case, it is possible to determine whether or not the purchased product is regularly changed for a certain item from the information of the purchase history, purchase cycle, and purchase change history for each user. If the purchased product is changed regularly, a product that the user has not purchased in the same item is recommended. For example, if another product is purchased each time for PET bottle tea, a product that has not been purchased is recommended. Such a user who regularly changes to another product is considered to be a user who has a desire to enjoy various products. According to this information recommendation device, it is possible to provide effective information for such a user.
[0024]
The information of a newly recommended product may be provided near the next scheduled purchase date in accordance with the purchase cycle, or may be provided along with a bargain sale date or a release date of a new product.
[0025]
An information recommendation destination extracting apparatus according to a third aspect of the present invention provides an information recommendation apparatus that recommends information on a product suitable for a user based on a user's purchase history and purchase cycle information on a plurality of products classified for each item. A history storage unit for storing purchase history information and purchase cycle information of a certain user's product; and a purchase history information and purchase cycle information of a certain user's product stored in the history storage unit. Purchase change history extraction means for extracting a change history of a purchased product of a user; and a feature for extracting a feature of change of a predetermined parameter representing a feature of the product before and after the change of the purchased product of a user. Based on the extracting means and the characteristic of the change of the predetermined parameter extracted by the characteristic extracting means, the characteristic of the predetermined parameter change is And means for identifying a product showing a change of similar parameters, and means for recommending the user with information about the product which is the identified as.
[0026]
When there is a change in the product purchased by the user, the feature of parameter change between the product before the change and the product after the change is extracted, and information is recommended based on the feature. To explain with the example of milk given above, if the product before change is milk with no component adjustment and the product after change is low-fat milk, the fat parameter of the milk components before and after the change Is changing. From this information, it can be determined that the user has changed to a tendency to purchase one with a lower fat content. Therefore, low-fat products are recommended for other items. For example, in the case of cheese, we recommend cottage cheese, which is lower in fat than cheese.
[0027]
In this way, by evaluating the detailed characteristics of the product using parameters and paying attention to the change, it is possible to provide even a product that is difficult to attract the user's interest only from the product name according to the user's preference. In addition, it is possible to provide information corresponding to the time of change in preference.
[0028]
The information of a newly recommended product may be provided near the next scheduled purchase date in accordance with the purchase cycle, or may be provided along with a bargain sale date or a release date of a new product. To explain with the example of the cheese described above, information on the cottage cheese may be provided near the next scheduled purchase date from the cheese purchase cycle.
[0029]
Preferably, the information providing unit includes a unit for calculating a next scheduled purchase date of a certain user regarding the product determined by the product determination unit from the purchase cycle information of the certain user, and information regarding the product determined by the product determination unit. Recommendation information storage means for storing the next scheduled purchase date, and information stored in the recommended information storage means in response to a predetermined information recommendation request, and the next scheduled purchase date. Means for providing a certain user with a date satisfying a predetermined information recommendation request and satisfying a predetermined relationship.
[0030]
By providing the user with information that satisfies a predetermined relationship with the next scheduled purchase date, the user can effectively use the information on the next scheduled purchase date.
[0031]
An information recommendation destination extraction device according to a fourth aspect of the present invention is an information recommendation destination extraction device for extracting a user appropriate as a recommendation destination of predetermined information from a purchase history of a user regarding a product, A purchase history storage unit for storing purchase histories of products of a plurality of users, and a change history of purchase products of a plurality of users is extracted for each user from purchase histories of products of the plurality of users stored in the purchase history storage unit. And a user extracting unit for extracting, as a recommendation destination of information, a user whose change history characteristic for each user extracted by the purchase change history extracting unit satisfies a predetermined condition.
[0032]
By extracting users whose change history characteristics satisfy predetermined conditions, it is possible to provide effective and appropriate information for users who change purchased items according to such characteristics. For example, by using this information as an advertisement for a product, the effect of the advertisement can be enhanced.
[0033]
Preferably, the extraction means includes means for extracting a user whose number of changes of the purchased product satisfies a predetermined condition as a recommendation destination of the information, based on the change history extracted by the purchase change history extraction means.
[0034]
When the number of times of change of the purchased product satisfies a predetermined condition, it is possible to provide the user with information meeting the condition. For example, if the number of changes is large, providing information on a new product increases the possibility that the user will purchase the product. In the case of a user whose number of times of change is small, it is possible to prepare in advance the contents of information to be provided in consideration of the contents.
[0035]
More preferably, the extraction means provides the user with information on the number of purchases of the product less than a predetermined threshold after the purchase product was last changed based on the change history extracted by the purchase change history extraction device. Including means for extracting as the first.
[0036]
A user who has made a small number of purchases after changing the purchased product is more likely to change the purchased product than other users. Therefore, by providing information to such a user, the information is used more effectively.
[0037]
An information recommendation apparatus according to a fifth aspect of the present invention recommends information to a user extracted by the information recommendation destination extraction apparatus in response to the information recommendation destination extraction apparatus and a predetermined information recommendation request. Means for
[0038]
A computer program according to a sixth aspect of the present invention, when executed by a computer, causes the computer to operate as the information recommendation device or the information recommendation destination extraction device described above.
[0039]
A computer-readable recording medium according to a seventh aspect of the present invention stores the computer program according to the fourth aspect.
[0040]
BEST MODE FOR CARRYING OUT THE INVENTION
Embodiments of the present invention will be described below with reference to the drawings. It should be noted that, in the present specification, the term “product” generally refers to an object of a transaction whether tangible or intangible. Generally, in the case of a transaction, a transfer of consideration such as money occurs in connection with the transaction. However, in the present specification, "goods" include not only such a transfer of money but also a transfer of money. It is a concept that also includes things.
[0041]
FIG. 1 is a block diagram showing the structure of the information recommendation device according to the present embodiment. With reference to FIG. 1, this information recommendation device can communicate with a portable information processing device such as a mobile phone or a personal digital assistant (hereinafter, referred to as “PDA”) owned by many users via a network. A possible device, specifically realized by a computer and its peripheral devices, and a program executed on the computer.
[0042]
Referring to FIG. 1, the information recommendation device receives an input of user identification information and purchase information by an operator, and further includes an input unit 1 for receiving a recommendation request, a CPU (Central Processing Unit) 2, and a CPU 2. A memory 3 that is connected and loaded with a program to be executed by the CPU 2 at the time of execution; a program storage unit 4 for storing a program for realizing an information recommendation device described later that is loaded into the memory 3 and executed by the CPU 2; It includes a history table 5, a purchase cycle table 6, a purchase change history table 7, and a recommended product list 8. Details of the configuration of each table will be described later.
[0043]
The information recommendation device further includes a transmission unit 9 for transmitting information about the recommended product determined by the information recommendation program to a corresponding user via a network, and a display unit 10 for displaying information about the recommended product. Including. Note that only one of the transmission unit 9 and the display unit 10 may be provided.
[0044]
Hereinafter, the purchase history table 5, the purchase cycle table 6, the purchase change history table 7, and the recommended product list 8 will be described. In the following description, it is assumed that the purchase situation of the user is an example as shown in FIG.
[0045]
FIGS. 2A and 2B show the purchase status of yogurt by user X and user Y in chronological order. The user X starts purchasing the product A on January 1 and thereafter purchases the same product every five days until December 16. The user Y also starts purchasing the product A on January 1 and has purchased the same product every five days until December 9, but has purchased the product B after December 6. The purchase interval of the product B is also every five days.
[0046]
FIG. 3 shows an example of the purchase history table 5. The purchase history table is a table that records a product name, a purchase date, and a purchase interval for each user and for each item of the purchased product. The purchase history table 5 is updated each time the user makes a purchase, and records purchase information on all purchased items. Search for the most recent purchase date with the same item name, and record the number of days between purchase dates as a purchase interval.
[0047]
FIG. 4 shows an example of the purchase cycle table 6. The purchase cycle table records the number of purchases and the average purchase cycle for each user and each purchased product. The purchase cycle table 6 is also updated each time the user makes a purchase. That is, for all the products purchased by the user from the purchase history table, the purchase interval from the previous purchase date is read, the average purchase cycle is calculated, and the information in the purchase cycle table 6 is updated. In addition, the number of purchases is incremented each time an update is performed.
[0048]
When the content of the purchase cycle table 6 is as shown in FIG. 4, the number of purchases of the product A of the user X is 71 times, the average purchase cycle is 5 days, and the number of purchases of the product A of the user Y is Is 70 times, the number of purchases of the product B is 73 times, and the average purchase cycle is 5 days.
[0049]
FIG. 5 shows an example of the purchase change history table 7. The purchase change history table is for recording a change history of purchased items of each item for each user. In the present embodiment, the purchase change history table 7 stores, when each user changes a product for a certain product, the product name, the product name before and after the change, the average purchase cycle of the product before the change, It records the number of purchases of the product before and after the change. The purchase change history table 7 is updated when each user purchases an item different from the previously purchased item for a certain item. In the purchase change history table, when a user first changes a purchased product for a certain item, an entry for the user for that item is created and stored. After the history is stored in the purchase change history table, the number of purchases after the change is updated each time the product after the change is purchased.
[0050]
FIG. 6 shows an example of the recommended product list 8. The recommended product list exists for each user, and stores an item of a product to be recommended to the user, a product name, and a planned purchase date.
[0051]
These tables are updated and referenced in the flow of processing described below. First, a basic idea of how these tables are updated based on a specific example will be described. Thereafter, a flowchart of a program for updating and recommending these tables and realizing an update / recommendation process will be described.
[0052]
Furthermore, in the present embodiment, when recommending product information to a user, the change history of another user is used. Hereinafter, the contents will be described.
[0053]
As a specific example, it is assumed that the user X regularly purchases the product A in the yogurt, but recently the purchase interval is long. That is, as shown in FIG. 2A, the user X purchases the product A in a cycle of 5 days, and should have purchased it on the 355th day after the 350th day. Even after that, it may not be purchased. User X was tired of product A because it had not been purchased even when the purchase time had come, although it had been purchased regularly in the past. It is presumed that it is included.
[0054]
At this time, attention is paid to another user Y as shown in FIG. Then, from FIG. 2B, it can be seen that the user Y has periodically purchased the product A for a long period of time, like the user X, but has changed the purchased product from the product A to the product B at a certain time. . Thus, from the past purchase history, if there is a user Y who has purchased the same product A as the user X in the same purchase cycle and has recently switched from the product A to another product B, the user Y Recommends to the user X the product B that has been newly purchased.
[0055]
For that purpose, the following processing is performed. From the purchase history table 5 and the purchase cycle table 4, it is determined that the user X has regularly purchased the product A of yogurt, and that the user X has purchased the product A despite an interval from the last purchase date. I understand that there is no. At first, since the user X may have forgotten to buy, the user X may be informed that the user A has forgotten to buy the product A. If the user X still does not purchase the product A, it can be determined that the user X is tired of the product A. In this case, the user X is notified of information on a new product different from the product A for this item.
[0056]
In order to provide new product information to the user X, the purchase change history table 7 is used. A search is made in the purchase change history table 7 to extract a change history of the user whose purchase frequency and purchase cycle of the product A are similar to the user X. Further, a recommended product for the user X is determined using information on what kind of purchasing behavior the user thus extracted has taken on this item.
[0057]
In the example illustrated in FIG. 5, the purchased product before the change is the product A, and the user Y whose purchase count and purchase cycle are similar to the user X can be determined to be similar users. In the example shown in FIG. 5, the user Y has changed the purchased product from the product A to the product B for the item called yogurt. Therefore, in this case, the product B is recommended to the user X as a new product.
[0058]
By using the purchase change history of the similar user in this way, it is possible to refer to which product of the same item has been changed by the user Y who likewise liked the product purchased regularly by the user X. Can be. Both are very similar in that user X and user Y liked the same product, and both have stopped purchasing product A. Therefore, it is considered that the product (product B) after the user Y changes the purchased product easily matches the preference of the user X. Therefore, when this information is presented to the user X, the possibility that the user X is interested in the product B is high, and the user X can be provided with useful information.
[0059]
Note that, in the present embodiment, both the number of purchases and the purchase cycle are similar as criteria for determining similar users. However, the present invention is not limited to such an embodiment. For example, a condition that at least one of them is similar may be used. Also, when a numerical value is used as a similar reference, the numerical values do not have to exactly match. For example, when two users have purchased a certain product at least a predetermined number of times, it may be determined that the two users are similar, or the purchase cycle of the product with the two users may both be a predetermined number of days. In the following cases, both users may be determined to be similar. As a result, a person who has made a regular purchase at a certain cycle is determined to be a similar user.
[0060]
One or more similar users may be extracted in the above processing. When a plurality of similar users are extracted, the changed product of the user that is considered to be most similar to the target user may be recommended, or the changed products of the extracted plurality of similar users may be recommended. The most popular products may be recommended.
[0061]
Further, in determining whether or not the user is a similar user, the number of purchases of the changed product may be used instead of the number of purchases and the purchase cycle of the product before the change. For example, the user W in FIG. 5 has a small number of purchases of the product A before the change, but has a large number of purchases of the product C after the change. Accordingly, it is expected that the product C can satisfy the user more than the product A. Therefore, in this case, the product C may be recommended.
[0062]
When selecting a recommended product from the purchase change history of a similar user, it may be determined whether the changed product is a temporarily changed product and selected. If the number of purchases of the changed product is only one, or if the number of purchases of the changed product is small and the product is changed to another product, the change is a temporary change. Therefore, in such a case, the change history may not be referred to.
[0063]
After determining a recommended product, the product and the next purchase date calculated from the purchase cycle are registered in the recommended product list shown in FIG. Then, when a date having a predetermined relation with the next scheduled purchase date, such as near the next scheduled purchase date, for example, one week before or next day of the next scheduled purchase date, is recommended to the user. However, the estimated purchase date may not be calculated. For example, the user X shown in FIG. 3 has not purchased yogurt after December 16. Therefore, in this case, a new product is recommended immediately, the planned purchase date is not calculated, and the recommended purchase list 8 does not record the planned purchase date. Then, among the entries in the recommended product list 8, those for which the planned purchase date is not recorded are immediately recommended to the user.
[0064]
As described above, when it is determined from the purchase history and the purchase cycle information that the user is tired of a certain product, a new product can be recommended from the change history of another similar user. With this method, it is possible to provide new information meaningful to the user at an appropriate timing.
[0065]
It should be noted that there is a user's purchase interval, and when deciding whether or not it is time to change the purchase, the user may simply forget to purchase or may postpone the purchase and purchase it on a cheap day. Therefore, instead of simply determining whether or not the purchase interval exceeds a certain threshold, the variance of the average purchase cycle may be determined, and the determination may be made in consideration of the variance.
[0066]
Further, in the present embodiment, a process of recommending another product of the same item to a certain user is also performed. Hereinafter, the details will be described.
[0067]
FIG. 7 shows an example of the purchase history of tea in the plastic bottle of the user X. FIG. 7 shows that the user X changes the purchased product every two times. That is, the user X changes to another product after purchasing the same product twice. By reading such a pattern from the purchase history status, after the user X purchases a certain product twice, when the next scheduled purchase date approaches, another product is selected from the unpurchased products and recommended. Can be adopted. By presenting such information to the user X, the user X can receive appropriate information that matches his / her own tea purchase pattern at an appropriate time.
[0068]
Therefore, in the present embodiment, processing is performed based on the following concept. The purchase change history table 7 as shown in FIG. 8 is obtained from the tea purchase status of the user X shown in FIG. From the purchased product change history table 7 in FIG. 8, it is determined whether or not this user frequently changes purchased products for a certain item. As shown in FIG. 8, when the number of purchases before and after a change is small for a certain product and the number of changes is large, it is determined that the product is frequently changed. Therefore, in this case, the merchandise O that has not been purchased by the user is registered in the recommended merchandise list in FIG. 6, and the information on the merchandise O is provided before the next scheduled tea purchase date.
[0069]
As described above, if the timing of switching the product by the user based on the purchase history information of the user is known, new information may be presented to the user at a time when a change is expected, and the new information may be provided to the user at an appropriate timing. Can be received at
[0070]
The product recommended here may be any product as long as it is an unpurchased product of the same item. Of course, the product may be determined based on other information. For example, a product, a new product, or the like, which is scheduled for sale near the scheduled purchase date, may be preferentially recommended.
[0071]
Further, as one feature of the present embodiment, when the user changes the purchased product, a process of recommending another product is performed based on the feature of parameter change of the product before and after the change. Hereinafter, the details will be described.
[0072]
10 shows an example of the user's milk and cheese purchase history status shown in FIG. 9. This user has changed from milk A, which has been purchased so far, to milk B, which is low-fat milk. By changing to low-fat milk, it can be seen that this user is trying to save fat. In such a case, it is desirable to recommend a low-fat product for another item, for example, an item such as cheese.
[0073]
FIG. 10 shows examples of components of milk and cheese for each product. In the example shown in FIG. 10, it can be seen that the amount of lipid is significantly different between product A and product B. Therefore, when the user changes the milk to be purchased from the product A to the product B, as the cheese to be recommended to the user, the product Y having a lower fat than other products is recommended as in the case of the product B. In this way, the various parameters of the product before and after the change are compared, and if the change has a remarkable feature, the change parameter is focused on and the product suitable for the change is extracted.
[0074]
If there is a product extracted by the above process, as shown in FIG. 11, the product Y is registered in the recommended product list 8 of the user, and the planned date of the next cheese purchase is stored. Then, it is recommended that the cheese be changed from the product X to the product Y before the scheduled purchase date.
[0075]
By evaluating a product using parameters as described above, it is possible to read a change in the user's preference from a user's purchase change and recommend a product according to the change in the preference. Any parameters can be evaluated. For example, description of a composition table, price, company name, material name, and the like can be considered.
[0076]
In the example given here, it is proposed to change the product within the item of cheese that the user has purchased regularly. On the other hand, it is also possible to recommend a product of an item that the user has not purchased before.
[0077]
For example, it is assumed that a certain user changed the diet method to a low insulin diet, and thus did not purchase meat or fish until now, but now purchases meat and conversely does not purchase rice. In this case, the user compares the parameters of the item that the user no longer buys and the item that the user buys, and if it is determined that the user only purchases the item having a low GI value (glycemic index index) parameter, It is determined that the eating tendency has changed to a preference for one having a low GI value, and an item having a low GI value is recommended as another item. The GI value is also referred to as a blood sugar increase reaction value, and it is said that a food having a higher value is a food that is more likely to be fat.
[0078]
As described above, by reading the change in the tendency of the user to purchase the product from the change in the parameter, the user can not only propose a change in the product of the item that the user has purchased regularly until now, but also purchase the product less frequently. Items that have not been made can be recommended.
[0079]
Hereinafter, a processing flow for implementing the above-described processing will be described.
[0080]
FIG. 12 is a flowchart showing the flow of processing of a program for realizing the information recommendation device according to the present invention. Referring to FIG. 12, it is determined whether or not the user has made a purchase (100). When the user has made a purchase and the purchase information is input (the determination result in step 100 is YES), it is determined in step 102 whether information updating and recommendation processing (step 104) has been performed for all the products. If the determination result is NO, information update and recommendation processing are executed in step 104, and control is returned to step 102 again. If it is determined in step 102 that information updating and recommendation processing have been completed for all products, control proceeds to step 106.
[0081]
In step 106, it is determined whether or not there is a product with a purchase interval for the user. If there is no such product, the process ends. If there are products with purchase intervals, it is determined in step 108 whether or not recommendation of information using the purchase history of another user has been performed for all such products. If the information recommendation has been completed for all, the process is terminated. If there is a product that has not been processed among the products with an interval, recommendation processing of information using the history of another user is performed on the product in step 110, and the control returns to step 108 again.
[0082]
In the following, details of the information update / recommendation process in step 104 shown in FIG. 12 and details of the information recommendation process using the history of another user in step 110 will be described.
[0083]
FIG. 13 shows a flowchart of a program for realizing the information update / recommendation process performed in step 104 of FIG. Referring to FIG. 13, first, at step 120, one purchased product is selected. In step 122, the purchase history table is updated for this user's item of this user (122), and the purchase cycle table is further updated (124).
[0084]
Subsequently, in step 126, it is determined whether the user has previously purchased the item. If it has been purchased before, control proceeds to step 132; otherwise, control proceeds to step 128.
[0085]
In step 132, it is further determined whether or not the purchased product this time is the same as the previously purchased product. In the same case,In step 133, in the purchase change history table, search for an entry for this user in which the product purchased this time is “product after change”, and add “1” to “number of purchases of product after”.This processing ends. As a result, the control returns to step 102 in FIG. 12, and proceeds to the processing of the next purchased product.
[0086]
If it is determined in step 132 that the purchased product is different from the previous purchased product, the purchase change history table is updated in step 134. That is, as shown in FIG. 5, for the item of the product of the user, the last purchased product, the current purchased product, the average purchase cycle of the previous product, and the number of purchases of the previous product are updated. . Further, “1” is recorded as the number of purchases of the subsequent product. Thereafter, this processing ends.
[0087]
On the other hand, if it is determined in step 126 that the user has not purchased the item of the product before, the purchase change history table is updated in step 128. Here, for the combination of the user and the item, a new entry including the changed product and the number of purchases ("1") is created and stored in the purchase change history table.
[0088]
Subsequently, control proceeds to step 130. In step 130, it is determined whether or not the product purchased this time is a product previously recommended by the present apparatus. If the product purchased by the user this time is a product recommended by the present apparatus, the process ends and the process proceeds to the next purchased product. If not, the control proceeds to step 136.
[0089]
In step 136, there is a periodicity in the change of the purchased product with respect to the item of the product purchased this time.If the next purchase follows the cycle, it corresponds to changing the purchased product.Is determined. If the user has periodically changed the purchased item in a certain cycle including a short period such as once or twice for the item, search for an unpurchased item in the same item (138). If there is a product (144), it is registered in the recommended product list (146). If it does not exist, do not register.
[0090]
If it is determined in step 136 that the purchased product has not been changed for the item, or if the product before the change has been purchased for a long time, in step 140, the characteristic parameters of the purchased product before and after the change are changed. Determine if there is.Here, for example, the amount of change is calculated for each of the characteristic parameters of the purchased product before and after the change, and it is checked whether or not there is a characteristic parameter indicating the amount of change exceeding a predetermined threshold. When there are a plurality of such parameters, a parameter having the largest change amount may be selected, or a combination of all or some of the plurality of parameters may be used in a later search for a product. good.If there is no change, this process ends. As a result, the process proceeds to the next purchase product.
[0091]
If it is determined in step 140 that there is a change in the parameter, a product matching the change in the parameter is searched for in another item (142). If such a product exists (YES in step 144), a recommendation is made. Register it in the product list (146). If not, the process proceeds to the next purchased product.In the search in step 142, for the user, among the products other than the purchased products of the other items, those showing the parameter change having the same tendency as the above-mentioned parameter change between the user and the purchased products of the item are searched. I do.
[0092]
FIG. 14 is a flowchart of a program that implements the process of making a recommendation using the history of another user, which is performed in step 110 of FIG. Referring to FIG. 14, when there is a product with a purchase interval, one product is selected (160), and it is determined whether or not the product was recommended to the user as a product with a previous interval (162). If no recommendation has been made before, the product is registered in the recommended product list because the user may have simply forgotten to buy it (S170). By registering in the recommended product list, the product is recommended to the user the next time the user makes a request for recommending the product, and the user is notified that he / she has forgotten to buy the product. In this case, the process ends when registered in the recommended product list.
[0093]
If it is determined in step 162 that the item is a previously recommended product, control proceeds to step 164. When the control proceeds to step 164, it means that the user has not purchased the product despite recommending it before, and it is time to change the purchase. Therefore, it is determined that a different product is recommended to this user, and the process from step 164 is performed.
[0094]
First, in step 164, a similar user is searched based on the purchase cycle table 6. At step 166, it is determined whether a similar user exists. If there is a similar user, the changed product is extracted from the purchase change history table of the similar user (168) and registered in the recommended product list 8 (170). If there is no similar user, the product is registered in the recommended product list 8 in step 170 in order to recommend the same product as the previous one to the user again.
[0095]
Returning to FIG. 12, when the processing in step 110 is performed for all the products with intervals, the determination result in step 108 becomes YES, and the entire processing ends.
[0096]
By repeating the above processing every time the user purchases a product, the recommended product list for the user in the recommended product list 8 is always kept up to date. Therefore, when the user requests the product recommendation process, the products stored in the recommended product list 8 may be presented. In this case, if there is a request, it is possible to present a product immediately based on the contents of the recommended product list 8, and the process does not require time.
[0097]
The flow of processing at the time of information recommendation will be described. FIG. 15 shows a flowchart of the process at the time of information recommendation. In this system, it is assumed that a user carries a mobile phone, a PDA, an IC (integrated circuit) card, or the like as a terminal, and makes a request for information recommendation to the terminal immediately before or immediately after a cashier clears or at the next shopping. are doing. In particular, information recommendation immediately before or immediately after cashier settlement is effective for notifying a product that has been forgotten to be purchased.
[0098]
In this embodiment, the recommendation request assumes that the user makes a recommendation request directly to the terminal. However, the present invention is not limited to such a case. For example, by mounting a GPS (Global Positioning System) function on a terminal, the terminal detects the position of the user, and when it is determined that the user is at a specific position, the device automatically issues a recommendation request. Is also good.
[0099]
Referring to FIG. 15, in step 180, it is determined whether or not there is a recommendation request, and the process stands by until there is a recommendation request. If there is a recommendation request (the determination result of step 180 is YES), a recommended product is extracted from the recommended product list 8 in step 182. Here, it is assumed that a product whose purchase date is close is preferentially extracted as a recommended product, and in the case immediately before and immediately after cashier settlement, a product with a purchase interval is preferentially extracted.
[0100]
In step 184, the information of the extracted recommended merchandise is provided to the user, and the process ends. That is, recommended products are displayed on the display screen provided in the terminal. When the recommendation process is automatically performed, the user may be notified by an alarm sound.
[0101]
As described above, according to the apparatus of the present embodiment, the products with purchase intervals are presented to the user as recommended products. For this reason, it is possible to prevent forgetting to buy a product. When there is a possibility that the user is going to purchase another product with the same item, an appropriate product is presented to the user at an appropriate time based on a similar purchase history of another user. The user can purchase a favorite product based on the information. In addition, if the user's purchase tendency changes, the user can recommend another item according to the change, and the user can purchase a product that matches his / her lifestyle. It is easier to do.
[0102]
In addition, in the above-described embodiment, as can be seen mainly from the example of food, information on tangible and paid products is recommended. However, the present invention is not limited to such an embodiment, and can be applied to intangible or free products, for example, contents such as text or music, or services.
[0103]
At present, advertisements are often provided on terminals for users having terminals such as mobile phones. The purchase history table 5 may be used for providing such an advertisement. For example, when an advertisement for a new product is issued, a user who frequently switches products with respect to the item to be advertised may be extracted from the purchase change history table 5 and the advertisement may be provided to the user preferentially.
[0104]
The above method of providing advertisements seems to be effective for the following reasons. That is, a user who has been buying a certain product for a long time is unlikely to make a replacement even if a new product comes out. On the other hand, a user who frequently switches products and tries various products is likely to react sensitively to an advertisement. Providing an advertisement to such a user can be expected to increase the advertisement effect.
[0105]
Also, it is considered that the advertisement providing effect is high even for a user who has purchased the changed product less frequently. This is because such a user is not familiar with the changed product, and it is considered that there is a high possibility that the user will buy a new product. Further, depending on the product, it may be more effective to provide an advertisement to a user who has been using the same product for a long time. The point is that it is possible to match the character of the product related to the advertisement with the tendency of the user to purchase the product.
[0106]
It is considered efficient to provide such advertisements at the same timing as the information recommendation shown in FIG. FIG. 16 shows a processing flow of providing an advertisement. First, at step 200, it is determined whether there is an advertisement request. In this embodiment, an advertisement request is made by a person who issues an advertisement (a provider of goods or the like). Therefore, it may be considered that the processing shown in FIG. 16 is executed independently of the processing shown in FIGS.
[0107]
If it is determined in step 200 that an advertisement providing request has been made, the user refers to the purchase history table 5 of the item of the product provided by the information recommendation and extracts a user having a large change history (202). There are various methods for extracting the user here. For example, a predetermined number of users who have the highest number of change histories may be extracted from all the users, or a user whose number of change histories is equal to or greater than a threshold may be extracted.
[0108]
Subsequently, in step 204, in addition to the users extracted in step 202, the purchase history table 5 is referenced to extract a user who has purchased the changed product less frequently. Here, users whose number of purchases of the changed product is equal to or less than a predetermined threshold value (different from the threshold value used in step 202) are extracted.
[0109]
Finally, in step 206, advertisement information is provided to the user extracted by the above processing. More specifically, the advertisement information list is stored in the same list as the recommended product list 8, and the advertisement is provided to the user at the same timing as the provision of the recommendation information to the user.
[0110]
Thus, by providing an advertisement to a user having a high advertisement effect, the advertisement cost is small and a high advertisement effect can be obtained.
[0111]
In the above embodiment, the information and the advertisement regarding the recommended product are provided using the terminal that can be carried by the user. However, the present invention is not limited to such an embodiment. For example, it is also possible to provide information using a non-carrying device such as a personal computer. In addition, if the user can identify the user by communicating with the IC card by possessing the IC card or the like, the recommendation information is displayed on the cashier display screen or the advertisement is presented at the checkout of the supermarket cashier. It is also possible to do. Furthermore, if a shopping basket used in a supermarket is provided with a communication function and a display function with an IC card and this device, recommendation information or an advertisement can be displayed on a display device of the shopping basket.
[0112]
The present invention has been described based on the embodiments. However, the embodiments of the present invention are not limited to the above-described embodiments, and various other modifications can be made. Further, it goes without saying that not only each element described in the above-described embodiment but also an embodiment that realizes a function equivalent to the function of the element and uses an alternative is included in the scope of the present invention.
[0113]
【The invention's effect】
As described above, according to the present invention, it is possible to provide effective information at a good timing for a user. Further, when information is provided using the change history of purchased products of similar users, it is possible to provide effective information to the original user when changing products. In addition, such information can be provided with good timing.
[0114]
If the user regularly changes the purchased product for an item, and recommends a product that the user has never purchased in the same item, the information provided to the user who takes such a purchase behavior is Can be used effectively.
[0115]
When there is a change in the product purchased by the user, the feature of parameter change between the product before the change and the product after the change is extracted, and information is recommended based on the feature. It is possible to provide even products that are difficult to pull according to the user's preference. In addition, it is possible to provide information corresponding to the time of change in preference.
[0116]
By providing the user with information that satisfies the predetermined relationship with the next scheduled purchase date, the user can effectively use the information at the next scheduled purchase date with good timing.
[0117]
If users whose change history characteristics satisfy predetermined conditions are extracted, it is possible to provide effective and appropriate information for users who change purchased goods according to such characteristics. For example, if the number of changes is large, providing information on a new product increases the possibility that the user will purchase the product. In the case of a user whose number of times of change is small, it is possible to prepare in advance the contents of information to be provided in consideration of the contents. As a result, it is possible to extract a target user who provides highly effective information at low cost.
[Brief description of the drawings]
FIG. 1 is a configuration diagram showing a structure of an information recommendation device according to an embodiment of the present invention.
FIG. 2 is a diagram illustrating a purchase situation of a user used in the embodiment of the present invention.
FIG. 3 is a diagram showing a configuration of a purchase history table of the embodiment of the information recommendation device according to the embodiment of the present invention.
FIG. 4 is a diagram showing a configuration of a purchase cycle table in the embodiment of the information recommendation device according to the embodiment of the present invention.
FIG. 5 is a diagram showing a configuration of a purchase change history table of the embodiment of the information recommendation device according to the embodiment of the present invention.
FIG. 6 is a diagram showing a configuration of a recommended product list of the embodiment of the information recommendation device according to the embodiment of the present invention.
FIG. 7 is a diagram illustrating a purchase situation of a user used in the embodiment of the present invention.
FIG. 8 is a diagram showing a configuration of a purchase change history table of the information recommendation device according to the embodiment of the present invention.
FIG. 9 is a diagram illustrating a purchase situation of a user used in the embodiment of the present invention.
FIG. 10 is a diagram showing parameters of a product of the information recommendation device according to the embodiment of the present invention.
FIG. 11 is a diagram showing a recommended product list used in the information recommendation device according to the embodiment of the present invention.
FIG. 12 is a flowchart showing an overall processing flow of a program for realizing the information recommendation device according to the embodiment of the present invention;
FIG. 13 is a flowchart showing a flow of information update / recommendation processing performed by the information recommendation device according to the embodiment of the present invention.
FIG. 14 is a flowchart showing a flow of information recommendation processing using a history of another user, which is performed by the information recommendation apparatus according to the embodiment of the present invention.
FIG. 15 is a flowchart of a flow of an information recommendation process performed by the information recommendation device according to the embodiment of the present invention.
FIG. 16 is a flowchart showing a flow of an advertisement providing process performed by the information recommendation device according to the embodiment of the present invention.
FIG. 17 is a diagram showing a purchase history of a user when explaining a conventional example.
[Explanation of symbols]
1 input unit, 2 CPU, 3 memory, 4 program storage unit, 5 purchase history table, 6 purchase cycle table, 7 purchase change history table, 8 recommended product list, 9 transmission unit, 10 display unit

Claims (14)

An information recommendation device for recommending information suitable for a user based on purchase history information and purchase cycle information of the user regarding products classified by items,
History storage means for storing purchase history information and purchase cycle information of a plurality of users;
A purchase change history extracting unit for extracting a change history of a purchased product of the user from the purchase history information and the purchase cycle information of the user stored in the history storage unit;
Determining means for determining whether it is time to change a purchased product for a certain item from the purchase history information and the purchase cycle information,
In response to the determining unit determining that the certain user is the time to change a purchased item for an item, based on the purchase change history of another user extracted by the purchase change history extracting unit, A product determination means for determining a product to be recommended to a user;
An information providing unit for providing information on the product determined by the product determining unit to the certain user. The product determination means,
A similar user extraction unit for searching the history storage unit and extracting a user who has a purchase tendency similar to the purchase tendency of the certain item by the certain user and has changed the purchase item of the certain item; ,
2. A means for determining, as a product to be recommended to the certain user, a changed purchased product of the certain item of the user extracted by the similar user extracting means. Information recommendation device. The similar user extracting means has a history of purchasing the same product as the purchased product of the certain item of the certain user at substantially the same cycle as the purchase cycle of the purchased product of the certain item of the certain user, 3. The information recommendation apparatus according to claim 2, further comprising means for extracting a user who has changed the purchased product. The information providing means,
Recommendation information storage means for storing information about the product determined by the product determination means,
2. The information recommendation apparatus according to claim 1, further comprising: means for providing information stored in said recommendation information storage means to said certain user in response to a predetermined information recommendation request. An information recommendation device that recommends information suitable for a user from purchase history information and purchase cycle information of the user regarding products classified by items,
History storage means for storing purchase history information and purchase cycle information of a user;
Purchase change history extraction means for extracting a change history of a purchased product from the purchase history information and purchase cycle information of the certain user stored in the history storage means,
From the purchase history information and the purchase cycle information, a determination unit for determining whether or not it is time for the user to change a purchased item for a certain item,
In response to the determination unit determining that the user is the time to change the purchased item for the certain item, recommending information on the unpurchased item of the user in the same item to the certain user. An information recommendation device including: an unpurchased product recommendation unit. An information recommendation device that recommends information on a product suitable for a user from a user's purchase history and purchase cycle information regarding a plurality of products classified for each item,
History storage means for storing purchase history information and purchase cycle information of a certain user's product,
Purchase change history extraction means for extracting a change history of the purchase product of the certain user from the purchase history information and purchase cycle information of the product of the certain user stored in the history storage unit,
A feature extraction unit for extracting a feature of a change of a predetermined parameter representing a feature of a product before and after a change of a purchased product of the certain user;
Based on the characteristic of the change of the predetermined parameter extracted by the characteristic extracting means, a product showing a parameter change similar to the characteristic of the change of the predetermined parameter for a product purchased by the certain user regarding another item. Means for identifying
Means for recommending information on the specified product to the certain user. The information providing means,
Means for calculating the next scheduled purchase date of the certain user for the product determined by the product determining means from the purchase cycle information of the certain user,
Information about the product determined by the product determination means, and recommendation information storage means for storing the next scheduled purchase date,
In response to the predetermined information recommendation request, the information stored in the recommendation information storage means and the next scheduled purchase date satisfies a predetermined relationship with the date on which the predetermined information recommendation request occurred. 4. The information recommendation apparatus according to claim 1, further comprising: means for providing a certain thing to the certain user. An information recommendation destination extraction device for extracting a user appropriate as a recommendation destination of predetermined information from a purchase history of a user regarding a product,
Purchase history storage means for storing purchase histories of products of a plurality of users;
From a purchase history of the products of the plurality of users stored in the purchase history storage unit, a purchase change history extraction unit for extracting a change history of the purchased products of the plurality of users for each user,
An information recommendation destination extraction device, comprising: a user extraction unit for extracting, as a recommendation destination of a user, a user whose change history characteristic for each user extracted by the purchase change history extraction unit satisfies a predetermined condition. 9. The information processing apparatus according to claim 8, wherein the extraction unit includes a unit for extracting a user whose number of changes of the purchased product satisfies a predetermined condition as a recommendation destination of information based on the change history extracted by the purchase change history extraction unit. The information recommendation destination extraction device described in 1. The extraction means, based on the change history extracted by the purchase change history extraction means, after the last change of the purchased product, the number of purchases of the product is less than a predetermined threshold user as a destination of information 9. The information recommendation destination extracting apparatus according to claim 8, further comprising an extracting unit. An information recommendation destination extraction device according to any one of claims 8 to 10,
Means for recommending information to a user extracted by the information recommendation destination extraction device in response to a predetermined information recommendation request. A computer program which, when executed by a computer, causes the computer to operate as the information recommendation device according to claim 1, claim 5, claim 6, or claim 11. A computer program that, when executed by a computer, causes the computer to operate as the information recommendation destination extraction device according to claim 8. A computer-readable recording medium on which the computer program according to claim 12 or 13 is recorded.

Images